This study was reviewed and approved by the Ethics Committee at the Chinese Academy of Medical Sciences and Peking Union Medical College Hospital (Beijing, China). We collected keloid scar specimens from 30 patients treated at Peking Union Medical College Hospital (Beijing, China). All patients provided written informed consent before participating in this study. All keloid cases were clinically and pathologically proven. Full-thickness skin specimens were harvested from the keloid scar tissues. No patients had previously received any treatment for keloids.

All skin specimens were incubated overnight at 4°C in dispase II (Sigma-Aldrich, St. Louis, MO, USA) to separated the epidermis from the dermis. Keratinocytes were extracted from the epidermis using 0.25% trypsin (Sigma-Aldrich) at 37°C for 30 min. All cells were filtered through a cell strainer (70 μm; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Primary keratinocytes were incubated in Keratinocyte-SFM (1X) (Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Invitrogen, San Diego, CA, USA) and 1X penicillin-streptomycin-fungizone (PSF; Life Technologies) for 48 or 72 h at 37°C to allow the cells to adhere to the culture dishes. Non-adherent cells were washed out with phosphate-buffered saline (PBS), and the remaining cells were subcultured or collected for the following analysis at approximately 80–90% confluence to avoid contact inhibition and differentiation. Cells at up to passage 3 were used for analyses.

In addition to the normoxic state, the influence of the hypoxic culture condition on epithelial cells was also determined. When the keratinocytes reached 50% confluence, they were then moved to a hypoxic incubation chamber (New Brunswick Scientific Co., Enfield, CT, USA). The hypoxic condition was maintained by continuous flushing with a mixed gas (1% O₂, 5% CO₂ and 94% N₂). The humidity was maintained, as well as the temperature (37°C).

The morphological alterations of the keratinocytes were observed by phase contrast microscopy. The keratinocytes were incubated under hypoxic conditions, and compared to the cells incubated under normoxic culture conditions for 12, 24 and 36 h. The morphological changes were recorded by using phase contrast microscopy and an Olympus inverted microscope (Olympus Corp., Tokyo, Japan).

Keloid biopsies were acquired and immediately fixed in 4% formaldehyde for 30 min. The specimens were then embedded in paraffin. Each slide was cut into 5-μm-thick sections. The slides were stained with different primary antibodies for immunohistochemistry, including mouse anti-E-cadherin (ab1416; 1:1,000), rabbit anti-vimentin (Ab92547; 1:1,000) (both from Abcam, Cambridge, UK), mouse anti-fibronectin (sc-8422; 1:400; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and mouse anti-HIF-1α antibodies (610958; 1:400; BD Biosciences, San Jose, CA, USA). All slides were incubated in biotin-labeled goat anti-rabbit (sc-45101) or goat anti-mouse (sc-395764) serum (1:200; Santa Cruz Biotechnology, Inc.) for 0.5 h, and detection was carried out using the VECTASTAIN Universal ABC-Alkaline Phosphatase kit with ImmPACT NovaRED Peroxidase Substrate (Vector Laboratories, Inc., Burlingame, CA, USA).

The expression of multiple secondary produced mesenchymal proteins, and HIF-1α protein in the epithelial cells was examined by immunofluorescence staining. We isolated primary keratinocytes as described above, then seeded them into dishes at a density of 1×10⁵ cells/well, and waited until they reached 70% confluence in each well under normoxic (21% O₂) or hypoxic (1% O₂) conditions. The cells were then fixed with 4% formaldehyde for 10 min. The cells were washed 3 times with PBS and the fixed cells were then treated with EMT-related and HIF-1α primary antibodies (described above) at the appropriate dilution at 4°C overnight, followed by incubation with FITC (sc-65218)- or TRITC (sc-2492)-conjugated secondary antibodies (1:150; Santa Cruz Biotechnology, Inc.) for 1 h at 37°C after extensively rinsing with PBS. Images were captured using an Olympus inverted fluorescence microscope (Olympus Corp.).

The siRNA expression vector for HIF-1α was constructed. To observe the HIF-1α-mediated changes in gene and protein expression under hypoxic conditions, we introduced the HIF-1α siRNA vector, pSilenser 2.1/HIF-si, including a HIF-1α-specific targeting sequence (Table I) and the control scramble siRNA vector, pSilenser, into the keloid keratinocytes. Cell transfection was performed using Lipofectamine 2000 (Life Technologies) according to the manufacturer’s instructions. The transfected cells were cultured for 48 h for further experiments under hypoxic conditions.

We used RT-qPCR to detect the expression of selected genes. Total RNA was extracted from the cultured keratinocytes derived from keloid scars using TRIzol reagent (Invitrogen), followed by treatment with DNase I (Promega Corp., Madison, WI, USA). First-stand cDNA was synthesized with 2 μg total RNA in 30 μl of reaction buffer using a high capacity cDNA synthesis kit (Takara Bio, Inc., Tokyo, Japan) according to the manufacturer’s instructions. Gene expression data were detected using the ABI StepOnePlus system (Applied Biosystems^®, Life Technologies). The thermal cycling parameters were 95°C for 1 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 40 sec. qPCR was performed to detect the mRNA levels using SYBR-Green I (Takara Bio, Inc.). The expression levels of genes were normalized to β-actin housekeeping gene expression. All RT-qPCR reactions were performed in technical triplicates. The primer sequences used for amplication are listed in Table I.

The protein levels of epithelial and mesenchymal markers in the keratinocytes were quantified by western blot analysis. Total protein was extracted using RIPA lysis buffer (Beyotime, Shanghai, China) which included complete Protease Inhibitor Cocktail (Sigma-Aldrich) and supernatants were collected by centrifugation at 13,000 × g at 4°C for 30 min. The protein concentration was detected using a QuantiPro BCA assay kit (Sigma-Aldrich) followed by separation on a 10% sodium dodecyl sulfate polyacrylamide gel. The proteins were relocated to 0.2 μm polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The proteins were blocked with Tris-buffered saline (TBS) with 5% skim milk for 1 h at room temperature. The membranes were then incubated with primary antibodies overnight at 4°C. Mouse anti-E-cadherin (ab1416; 1:1,000; Abcam), rabbit anti-zonula occludens-1 (ZO-1; sc-10804) (1:400; Santa Cruz Biotechnology, Inc.), rabbit anti-vimentin (Ab92547; 1:1,000; Abcam), mouse anti-fibronectin (sc-8422; 1:400; Santa Cruz Biotechnology, Inc.), mouse anti-HIF-1α (610958; 1:400; BD Biosciences) and mouse anti-β-actin (1:4,000; Sigma-Aldrich) were used as primary antibodies. Horseradish peroxidase-labeled goat anti-mouse/rabbit IgG (Beyotime) was then added as the secondary antibody. Signals were visualized using an Immobilon Western Chemiluminescent HRP Substrate (WBKLS0100; Millipore) and image detection was performed using the ImageQuant LAS 4000 mini imaging system (GE Healthcare Bio-Sciences, Pittsburgh, PA, USA). The results were normalized to β-actin (Sigma-Aldrich) where appropriate.

The invasiveness of the keloid keratinocytes under normoxic and hypoxic conditions was measured by Transwell co-culture assay. The keratinocytes (2.0×10⁵) were seeded into the top chamber of 24-well plates (pore size, 8 μm; Corning, Inc., Corning, NY, USA) covered with Matrigel (Coating Matrix; Life Technologies) and incubated in full supplemented medium. The lower chambers were contained with chemoattractant (20% serum culture medium). The whole plates were placed into either a normoxic or hypoxic chamber for incubation 12, 24 and 36 h for observation. In addition, the keratinocytes transfected with HIF-1α siRNA, control scramble siRNA and the untransfected keratinocytes were also analyzed in triplicate wells. For those cells that did not go through the membranes, they were wiped out using a cotton swap. Those traversed cells were fixed in methanol and stained in Crystal violet. Phase contrast microscopy was used to record digital images. Three microscopic fields were recorded in each well. At the same time, the elution of Crystal violet was measured at 570 nM spectrophotometer absorbance.

Continuous data are expressed as the means ± standard deviation. For statistical analysis, the Student’s t-test was carried out using SPSS Statistics 19.0 software (SPSS, Inc., Chicago, IL, USA). A value of P<0.05 was considered to indicate a statistically significant difference.

To demonstrate EMT characteristics in the keloid scars, keloid scar biopsy specimens were examined by histological analysis as shown in Fig. 1. Among these specimens, the epithelial cells clearly underwent EMT, as shown by the abnormal expression of the mesenchymal markers, vimentin and fibronectin, at the basement membrane zone of the epithelial layer, which is the migration zone to the dermis. However, epithelial markers, such as E-cadherin also exhibited a slightly decreased expression around the basement membrane area, which confirmed the occurence of transformation. In addition, HIF-1α was also stained in the same specimens, which coincidentally had a high expression at the same location as the mesenchymal markers (Fig. 1).

Morphometric analysis was performed to further estimate the degree of EMT induction under hypoxic culture conditions. The cells were observed at 12, 24 and 36 h. The keratinocytes cultured under normoxic conditions displayed a characteristic polygonal and cobblestone monolayer morphology. Following culture for 12 h under hypoxic conditions, phenotypic changes, from trivial to remarkable, were observed. Compared to the control cells (cultured under normoxic conditions), the hypoxia-stimulated keratinocytes displayed an elongated, spindle-shaped similar morphology, resembling fibroblasts (Fig. 2).

The induction of the expression of EMT marker genes was measured under both normoxic and hypoxic conditions. Following incubation for 12, 24 and 36 h in 1% oxygen, the mRNA level of HIF-1α increased by 3.5-, 7.5- and 10-fold, respectively (P<0.05, 0.05 and 0.05, respectively) compared to the controls. Accompanied by the increase in HIF-1α mRNA expression, the mRNA expression of vimentin increased by 3.4-, 3.6- and 11-fold, respectively (P<0.05, 0.05 and 0.05, respectively) compared to the controls. The mRNA expression of fibronectin increased by 10.5-, 12- and 17-fold, respectively (P<0.05, 0.05, and 0.05, respectively) compared to the controls. However, the mRNA expression of ZO-1 was suppressed under the hypoxic conditions by approximately 74.8±1.7% at 12 h, 69.0±3.3% at 24 h and 84.2±2.7% at 36 h compared to the controls (P<0.05, 0.05 and 0.05, respectively). The mRNA expression of E-cadherin in response to the hypoxic conditions decreased by approximately 5.8±1.3% at 12 h, 68.7±5.1% at 24 h and 72.6±2.3% at 36 h compared to the controls (P=0.30, P<0.05 and 0.05, respectively) (Fig. 3).

Due to the association between HIF-1α and EMT markers in keloid scar specimens and hypoxia-induced phenotypic changes in keratinocytes, HIF-1α may potentially provoke or intensify EMT markers during human keloid scar formation. To examine this hypothesis, we first cultured primary keloid-derived keratinocytes in a hypoxic environment (1% O₂). At 12, 24 and 36 h, the keratinocytes were collected for western blot analysis. In the normoxic control group, vimentin-and fibronectin-positive cells were absent, as shown by western blot analysis. Following culture under hypoxic conditions, the protein epxression of vimentin and fibronectin increased significantly (P<0.05 and 0.05, respectively). By contrast, the epxression of epithelial markers, including E-cadherin and ZO-1 was significantly decreased compared to the normoxic controls (P<0.05 and 0.05, respectively) (Fig. 4A).

To further demonstrate the expression of EMT markers in the cells cultured under normoxic and hypoxic conditions, immunofluorescence staining was performed to evaluate the expression of vimentin, fibronectin, E-cadherin, ZO-1 and HIF-1α. E-cadherin and ZO-1 were strongly manifested in the control group; no expression of mesenchymal markers was observed in the controls (Fig. 4B). However, following exposure to hypoxia, the expression of ZO-1 and E-cadherin was significantly downregulated, and that of mesenchymal markers was increased (Fig. 4C).

In order to confirm the mechanisms through which HIF-1α participates in hypoxia-mediated EMT in keloid-derived keratinocytes, the HIF-1α gene was knocked down by siRNA in keloid-derived keratinocytes. The expression levels of E-cadherin, ZO-1, fibronectin and vimentin were measured by western blot analysis following the silencing of HIF-1α by introducing HIF-1α siRNA or control scramble siRNA into the keratinocytes under hypoxic (1% O₂) culture conditions. A significant decrease in the expression of vimentin and fibronectin was observed (P<0.05 and 0.05, respectively). On the contrary, the marked restoration in the expression of E-cadherin and ZO-1 suggested the distinctive importance of HIF-1α signaling in EMT in keloid keratinocytes (P<0.05 and 0.05, respectively) (Fig. 5). The effects of siRNA on keloid keratinocytes under hypoxic conditions were also confirmed by phase contrast microscopy. The cellular morphology of the keratinocytes transfected with HIF-1α siRNA was reversed back to a typical epithelial-like shape compared to the control cells transfected with the scramble siRNA under hypoxic conditions (Fig. 6). This indicated that the mesenchymal changes induced by hypoxia in the keloid keratinocytes were regulated by HIF-1α signaling.

The stimulating effect of hypoxia on the invasion of keloid-derived keratinocytes was investigated under hypoxic culture conditions. The keratinocytes were cultured under hypoxic conditions (1% O₂) for 12, 24 and 36 h, and compared to the cells cultured under normoxic conditions. The number of migrated keratinocytes at 12, 24, and 36 h of culture under hypoxic conditions markedly increased by 9-, 15.8- and 32-fold, respectively compared to the normoxic controls (P<0.01, 0.01 and 0.01, respectively) (Fig. 7). In addition, the invasive capacity of the cells was compared between the HIF-1α siRNA-transfected keratinocytes and the hypoxia-exposed keratinocytes. The results revealed a significant decrease in the number of infiltrated cells by 91.3±1.2% in the HIF-1α siRNA-transfected group (P<0.05) (Fig. 8). As a result, hypoxia/HIF-1α may enhance the invasive capacity of keloid keratinocytes.